From: James Bottomley <James.Bottomley@HansenPartnership.com> To: Russell King - ARM Linux <linux@arm.linux.org.uk> Cc: "Trond Myklebust" <Trond.Myklebust@netapp.com>, "Linus Torvalds" <torvalds@linux-foundation.org>, linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org, "Marc Kleine-Budde" <mkl@pengutronix.de>, "Uwe Kleine-König" <u.kleine-koenig@pengutronix.de>, "Marc Kleine-Budde" <m.kleine-budde@pengutronix.de>, linux-arm-kernel@lists.infradead.org, "Parisc List" <linux-parisc@vger.kernel.org>, linux-arch@vger.kernel.org Subject: Re: still nfs problems [Was: Linux 2.6.37-rc8] Date: Thu, 06 Jan 2011 15:07:58 -0600 [thread overview] Message-ID: <1294348078.22825.346.camel@mulgrave.site> (raw) In-Reply-To: <1294338341.22825.216.camel@mulgrave.site> On Thu, 2011-01-06 at 12:25 -0600, James Bottomley wrote: > OK, so thinking about this, it seems that the only danger is actually > what NFS is doing: reading cache pages via a vmap. In that case, since > the requirement is to invalidate the vmap range to prepare for read, we > could have invalidate_kernel_vmap_range loop over the underlying pages > and flush them through the kernel alias if the architecture specific > flag indicates their contents might be dirty. > > The loop adds expense that is probably largely unnecessary to > invalidate_kernel_vmap_range() but the alternative is adding to the API > proliferation with something that only flushes the kernel pages if the > arch specific flag says they're dirty. This is what I think the arm patch would look like (example only: I can't compile it). Is something like this too expensive? the loop can't be optimised away because of the need to check the pages (and vmalloc_to_page is a three level page table lookup). James --- diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h index 3acd8fa..34469ca 100644 --- a/arch/arm/include/asm/cacheflush.h +++ b/arch/arm/include/asm/cacheflush.h @@ -414,8 +414,17 @@ static inline void flush_kernel_vmap_range(void *addr, int size) } static inline void invalidate_kernel_vmap_range(void *addr, int size) { - if ((cache_is_vivt() || cache_is_vipt_aliasing())) - __cpuc_flush_dcache_area(addr, (size_t)size); + if ((cache_is_vivt() || cache_is_vipt_aliasing())) { + void *cursor = addr; + + for ( ; cursor < addr + size; cursor += PAGE_SIZE) { + struct page *page = vmalloc_to_page(cursor); + + if (!test_and_set_bit(PG_dcache_clean, &page->flags)) + __flush_dcache_page(page_mapping(page), page); + } + __cpuc_flush_dcache_area(addr, (size_t)size); + } } #define ARCH_HAS_FLUSH_ANON_PAGE
WARNING: multiple messages have this Message-ID (diff)
From: James.Bottomley@HansenPartnership.com (James Bottomley) To: linux-arm-kernel@lists.infradead.org Subject: still nfs problems [Was: Linux 2.6.37-rc8] Date: Thu, 06 Jan 2011 15:07:58 -0600 [thread overview] Message-ID: <1294348078.22825.346.camel@mulgrave.site> (raw) In-Reply-To: <1294338341.22825.216.camel@mulgrave.site> On Thu, 2011-01-06 at 12:25 -0600, James Bottomley wrote: > OK, so thinking about this, it seems that the only danger is actually > what NFS is doing: reading cache pages via a vmap. In that case, since > the requirement is to invalidate the vmap range to prepare for read, we > could have invalidate_kernel_vmap_range loop over the underlying pages > and flush them through the kernel alias if the architecture specific > flag indicates their contents might be dirty. > > The loop adds expense that is probably largely unnecessary to > invalidate_kernel_vmap_range() but the alternative is adding to the API > proliferation with something that only flushes the kernel pages if the > arch specific flag says they're dirty. This is what I think the arm patch would look like (example only: I can't compile it). Is something like this too expensive? the loop can't be optimised away because of the need to check the pages (and vmalloc_to_page is a three level page table lookup). James --- diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h index 3acd8fa..34469ca 100644 --- a/arch/arm/include/asm/cacheflush.h +++ b/arch/arm/include/asm/cacheflush.h @@ -414,8 +414,17 @@ static inline void flush_kernel_vmap_range(void *addr, int size) } static inline void invalidate_kernel_vmap_range(void *addr, int size) { - if ((cache_is_vivt() || cache_is_vipt_aliasing())) - __cpuc_flush_dcache_area(addr, (size_t)size); + if ((cache_is_vivt() || cache_is_vipt_aliasing())) { + void *cursor = addr; + + for ( ; cursor < addr + size; cursor += PAGE_SIZE) { + struct page *page = vmalloc_to_page(cursor); + + if (!test_and_set_bit(PG_dcache_clean, &page->flags)) + __flush_dcache_page(page_mapping(page), page); + } + __cpuc_flush_dcache_area(addr, (size_t)size); + } } #define ARCH_HAS_FLUSH_ANON_PAGE
next prev parent reply other threads:[~2011-01-06 21:07 UTC|newest] Thread overview: 194+ messages / expand[flat|nested] mbox.gz Atom feed top 2011-01-05 19:05 still nfs problems [Was: Linux 2.6.37-rc8] James Bottomley 2011-01-05 19:05 ` James Bottomley 2011-01-05 19:18 ` Linus Torvalds 2011-01-05 19:18 ` Linus Torvalds [not found] ` <AANLkTi=VZUxNFd7n-qwf5aiOeK5rkk8qBmo+kOpgg7up-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> 2011-01-05 19:36 ` James Bottomley 2011-01-05 19:36 ` James Bottomley 2011-01-05 19:36 ` James Bottomley 2011-01-05 19:49 ` Linus Torvalds 2011-01-05 19:49 ` Linus Torvalds 2011-01-05 19:49 ` Linus Torvalds 2011-01-05 20:35 ` James Bottomley 2011-01-05 20:35 ` James Bottomley [not found] ` <1294256169.16957.18.camel-0iu6Cu4xQGLYCGPCin2YbQ@public.gmane.org> 2011-01-05 20:00 ` Russell King - ARM Linux 2011-01-05 20:00 ` Russell King - ARM Linux 2011-01-05 20:00 ` Russell King - ARM Linux 2011-01-05 20:33 ` James Bottomley 2011-01-05 20:33 ` James Bottomley 2011-01-05 20:48 ` Linus Torvalds 2011-01-05 20:48 ` Linus Torvalds 2011-01-05 20:48 ` Linus Torvalds [not found] ` <AANLkTimzzBsdtWcZtP5E_CH1hUZugGMoaHOiMdQJf764-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> 2011-01-05 21:04 ` Russell King - ARM Linux 2011-01-05 21:04 ` Russell King - ARM Linux 2011-01-05 21:04 ` Russell King - ARM Linux 2011-01-05 21:08 ` Linus Torvalds 2011-01-05 21:08 ` Linus Torvalds [not found] ` <AANLkTi=EXXBTW7oWHq3D+PHsx=thF1CpkRjn0ax2p5rm-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> 2011-01-05 21:16 ` Trond Myklebust 2011-01-05 21:16 ` Trond Myklebust 2011-01-05 21:16 ` Trond Myklebust [not found] ` <1294262208.2952.4.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> 2011-01-05 21:30 ` Linus Torvalds 2011-01-05 21:30 ` Linus Torvalds 2011-01-05 21:30 ` Linus Torvalds 2011-01-05 23:06 ` Trond Myklebust 2011-01-05 23:06 ` Trond Myklebust 2011-01-05 23:28 ` James Bottomley 2011-01-05 23:28 ` James Bottomley 2011-01-06 17:40 ` James Bottomley 2011-01-06 17:40 ` James Bottomley 2011-01-06 17:47 ` Trond Myklebust 2011-01-06 17:47 ` Trond Myklebust 2011-01-06 17:51 ` James Bottomley 2011-01-06 17:51 ` James Bottomley 2011-01-06 17:55 ` Linus Torvalds 2011-01-06 17:55 ` Linus Torvalds 2011-01-06 17:55 ` Linus Torvalds 2011-01-07 18:53 ` Trond Myklebust 2011-01-07 18:53 ` Trond Myklebust [not found] ` <1294426405.2929.23.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> 2011-01-07 19:02 ` Russell King - ARM Linux 2011-01-07 19:02 ` Russell King - ARM Linux 2011-01-07 19:02 ` Russell King - ARM Linux 2011-01-07 19:11 ` James Bottomley 2011-01-07 19:11 ` James Bottomley [not found] ` <1294427467.4895.66.camel-0iu6Cu4xQGLYCGPCin2YbQ@public.gmane.org> 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 16:49 ` Trond Myklebust 2011-01-08 23:15 ` Trond Myklebust 2011-01-08 23:15 ` Trond Myklebust 2011-01-08 23:15 ` Trond Myklebust 2011-01-08 23:15 ` Trond Myklebust [not found] ` <1294528551.4181.19.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> 2011-01-10 10:50 ` Uwe Kleine-König 2011-01-10 10:50 ` Uwe Kleine-König 2011-01-10 10:50 ` Uwe Kleine-König 2011-01-10 10:50 ` Uwe Kleine-König 2011-01-10 16:25 ` Trond Myklebust 2011-01-10 16:25 ` Trond Myklebust 2011-01-10 16:25 ` Trond Myklebust [not found] ` <1294676734.3349.10.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> 2011-01-10 17:08 ` Marc Kleine-Budde 2011-01-10 17:08 ` Marc Kleine-Budde 2011-01-10 17:08 ` Marc Kleine-Budde 2011-01-10 17:20 ` Trond Myklebust 2011-01-10 17:20 ` Trond Myklebust 2011-01-10 17:20 ` Trond Myklebust [not found] ` <1294680035.3349.19.camel-rJ7iovZKK19ZJLDQqaL3InhyD016LWXt@public.gmane.org> 2011-01-10 17:26 ` Marc Kleine-Budde 2011-01-10 17:26 ` Marc Kleine-Budde 2011-01-10 17:26 ` Marc Kleine-Budde 2011-01-10 17:26 ` Marc Kleine-Budde 2011-01-10 19:25 ` Uwe Kleine-König 2011-01-10 19:25 ` Uwe Kleine-König 2011-01-10 19:25 ` Uwe Kleine-König [not found] ` <20110110192552.GG24920-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org> 2011-01-10 19:29 ` Trond Myklebust 2011-01-10 19:29 ` Trond Myklebust 2011-01-10 19:29 ` Trond Myklebust 2011-01-10 19:29 ` Trond Myklebust 2011-01-10 19:31 ` James Bottomley 2011-01-10 19:31 ` James Bottomley 2011-01-10 19:34 ` Linus Torvalds 2011-01-10 19:34 ` Linus Torvalds 2011-01-10 20:15 ` Trond Myklebust 2011-01-10 20:15 ` Trond Myklebust 2011-01-10 12:44 ` Marc Kleine-Budde 2011-01-10 12:44 ` Marc Kleine-Budde 2011-01-10 12:44 ` Marc Kleine-Budde 2011-01-07 19:13 ` Trond Myklebust 2011-01-07 19:13 ` Trond Myklebust 2011-01-07 19:05 ` James Bottomley 2011-01-07 19:05 ` James Bottomley [not found] ` <1294335614.22825.154.camel-0iu6Cu4xQGLYCGPCin2YbQ@public.gmane.org> 2011-01-06 18:05 ` Russell King - ARM Linux 2011-01-06 18:05 ` Russell King - ARM Linux 2011-01-06 18:05 ` Russell King - ARM Linux 2011-01-06 18:14 ` James Bottomley 2011-01-06 18:14 ` James Bottomley [not found] ` <1294337670.22825.199.camel-0iu6Cu4xQGLYCGPCin2YbQ@public.gmane.org> 2011-01-06 18:25 ` James Bottomley 2011-01-06 18:25 ` James Bottomley 2011-01-06 18:25 ` James Bottomley 2011-01-06 21:07 ` James Bottomley [this message] 2011-01-06 21:07 ` James Bottomley 2011-01-06 20:19 ` John Stoffel 2011-01-06 20:19 ` John Stoffel 2011-01-05 23:28 ` Linus Torvalds 2011-01-05 23:28 ` Linus Torvalds [not found] ` <AANLkTi=SjMinMp+m726GS1iehj6cQgNy1RqSoUqKhjtv-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org> 2011-01-05 23:59 ` Russell King - ARM Linux 2011-01-05 23:59 ` Russell King - ARM Linux 2011-01-05 23:59 ` Russell King - ARM Linux 2011-01-05 23:59 ` Russell King - ARM Linux 2011-01-05 23:59 ` Russell King - ARM Linux 2011-01-05 21:16 ` James Bottomley 2011-01-05 21:16 ` James Bottomley -- strict thread matches above, loose matches on Subject: below -- 2010-12-31 3:17 George Spelvin 2010-12-31 4:32 ` Trond Myklebust 2011-01-01 1:03 ` George Spelvin 2011-01-01 1:18 ` Trond Myklebust 2011-01-01 5:44 ` George Spelvin 2010-12-29 1:18 Linux 2.6.37-rc8 Linus Torvalds 2010-12-30 17:14 ` still nfs problems [Was: Linux 2.6.37-rc8] Uwe Kleine-König 2010-12-30 17:14 ` Uwe Kleine-König 2010-12-30 17:57 ` Linus Torvalds 2010-12-30 17:57 ` Linus Torvalds 2010-12-30 18:24 ` Trond Myklebust 2010-12-30 18:24 ` Trond Myklebust 2010-12-30 18:50 ` Linus Torvalds 2010-12-30 18:50 ` Linus Torvalds 2010-12-30 19:25 ` Trond Myklebust 2010-12-30 19:25 ` Trond Myklebust 2010-12-30 20:02 ` Linus Torvalds 2010-12-30 20:02 ` Linus Torvalds 2010-12-30 17:59 ` Trond Myklebust 2010-12-30 17:59 ` Trond Myklebust 2010-12-30 19:18 ` Uwe Kleine-König 2010-12-30 19:18 ` Uwe Kleine-König 2011-01-03 21:38 ` Uwe Kleine-König 2011-01-03 21:38 ` Uwe Kleine-König 2011-01-04 0:22 ` Trond Myklebust 2011-01-04 0:22 ` Trond Myklebust 2011-01-05 8:40 ` Uwe Kleine-König 2011-01-05 8:40 ` Uwe Kleine-König 2011-01-05 11:05 ` Uwe Kleine-König 2011-01-05 11:05 ` Uwe Kleine-König 2011-01-05 11:27 ` Russell King - ARM Linux 2011-01-05 11:27 ` Russell King - ARM Linux 2011-01-05 12:14 ` Marc Kleine-Budde 2011-01-05 12:14 ` Marc Kleine-Budde 2011-01-05 13:02 ` Nori, Sekhar 2011-01-05 13:02 ` Nori, Sekhar 2011-01-05 15:34 ` Russell King - ARM Linux 2011-01-05 15:34 ` Russell King - ARM Linux 2011-01-05 13:40 ` Uwe Kleine-König 2011-01-05 13:40 ` Uwe Kleine-König 2011-01-05 14:29 ` Jim Rees 2011-01-05 14:29 ` Jim Rees 2011-01-05 14:42 ` Marc Kleine-Budde 2011-01-05 14:42 ` Marc Kleine-Budde 2011-01-05 15:38 ` Jim Rees 2011-01-05 15:38 ` Jim Rees 2011-01-05 14:53 ` Trond Myklebust 2011-01-05 14:53 ` Trond Myklebust 2011-01-05 15:01 ` Marc Kleine-Budde 2011-01-05 15:01 ` Marc Kleine-Budde 2011-01-05 15:14 ` Trond Myklebust 2011-01-05 15:14 ` Trond Myklebust 2011-01-05 15:29 ` Trond Myklebust 2011-01-05 15:29 ` Trond Myklebust 2011-01-05 15:39 ` Marc Kleine-Budde 2011-01-05 15:39 ` Marc Kleine-Budde 2011-01-05 15:52 ` Russell King - ARM Linux 2011-01-05 15:52 ` Russell King - ARM Linux 2011-01-05 17:17 ` Trond Myklebust 2011-01-05 17:17 ` Trond Myklebust 2011-01-05 17:26 ` Russell King - ARM Linux 2011-01-05 17:26 ` Russell King - ARM Linux 2011-01-05 18:12 ` Trond Myklebust 2011-01-05 18:12 ` Trond Myklebust 2011-01-05 18:27 ` Russell King - ARM Linux 2011-01-05 18:27 ` Russell King - ARM Linux 2011-01-05 18:55 ` Trond Myklebust 2011-01-05 18:55 ` Trond Myklebust 2011-01-05 19:07 ` Russell King - ARM Linux 2011-01-05 19:07 ` Russell King - ARM Linux 2011-01-14 2:25 ` Andy Isaacson 2011-01-14 2:25 ` Andy Isaacson 2011-01-14 2:40 ` Trond Myklebust 2011-01-14 2:40 ` Trond Myklebust 2011-01-14 4:22 ` Andy Isaacson 2011-01-14 4:22 ` Andy Isaacson
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=1294348078.22825.346.camel@mulgrave.site \ --to=james.bottomley@hansenpartnership.com \ --cc=Trond.Myklebust@netapp.com \ --cc=linux-arch@vger.kernel.org \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-nfs@vger.kernel.org \ --cc=linux-parisc@vger.kernel.org \ --cc=linux@arm.linux.org.uk \ --cc=m.kleine-budde@pengutronix.de \ --cc=mkl@pengutronix.de \ --cc=torvalds@linux-foundation.org \ --cc=u.kleine-koenig@pengutronix.de \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.